OBJECTIVES: We illustrate nine tests for assessing validity of a discrete-choice experiment (DCE). These test include examining simplifying heuristics, choice consistency, and face validity. METHODS: A paired-comparison DCE estimated treatment preference of patients with type 2 diabetes. We employed a blocked D-efficient experimental design. Results were analyzed using mixed-logit and latent class models. We assessed simplifying heuristic through frequency of task non-attendance, frequency of lexicographic preferences, and presence of attribute non-attendance. We assessed choice consistency through parallel-forms reliability across three survey versions, inter-choice consistency for a repeat task, and predictive accuracy for a holdout task. We assessed face validity by asking for participants’ evaluations, by evaluating direction of preferences, and by placing results in context of published preference studies. RESULTS: